scrap metal analysis

Intro

About

AI-powered scrap quality inspectionThe system uses artificial intelligence to assess scrap metal quality from photos. Operators review and confirm the results via desktop interfaces — this is a simplified overview of the process

Users

Scrap metal processing operators and yard operators

My role

Product Designer

Timeline

4 months

Problem & Goals

Problem

Operators have no way to control traffic lights from the system and are forced to call yard operators directly. They also cannot manually enter license plate numbers for vehicles that weren’t recognized by the system.

Goals

Design missing user screens to enable full workflow support, followed by interface analysis and field testing aimed at improving speed and efficiency for yard operators.

Constraints

As a development-driven product, designs couldn’t be implemented strictly according to testing results. Finding the right balance required close collaboration with the development team to propose feasible, high-quality solutions through compromise and iteration.

Discovery & Research

An analysis of 25 support tickets revealed the need for a redesigned main menu, as well as new screens for entry signal control and unloading status for both trucks and trains.

To clarify requirements, was conducted a usability test of the current main menu.
Key findings from the testing:
  • It was unclear which vehicles were trains and which were trucks
  • The «Open» and «Generate PDF report» buttons were rarely used from this screen
  • There was no way to manually input unrecognized license plate numbers, affecting reporting accuracy
  • Operators couldn’t determine the current location of a vehicle within the facility, causing frequent disruptions from calls between field and control room operators

As Is process

This is a simplified process as is. A lot of manual work and lack of control of the process

Approach & Process

Since I had UI freedom but several UX constraints in a collaboration with developers we started to create optimal solution

To Be Process

Less manual checking with yard operators
More control of the entering and unloading process
Greens is the new operator’s point of control

Following the design of the new screens, additional qualitative research was conducted to validate the effectiveness of the solution for operators.

Solution

The updated main menu now makes it easier to distinguish between trains and trucks, displaying vehicle numbers clearly. If a number is not recognized, it can be manually entered via a modal window.
Trains and trucks require different interfaces: trains are unloaded one car at the time, while trucks are unloaded all at once. Operators now manage both entry/exit and unloading processes using traffic light controls.
For the yard operators there is a tablet version. Main task for this users is to control process on-site and keep clear the information in the system.

Outcome & Impact

For off-site operators, system transparency significantly improved — they can now monitor traffic light status and more easily track incoming trucks and trains. Workflow speed increased by 20%, and onboarding time for new operators dropped by 40% thanks to the redesigned screens and updated main menu.
For the yard operators, the number of calls and messages decreased, allowing them to stay focused and make fewer mistakes. With the new screens, error rates were reduced by 14%.

Reflection

It was truly exciting to be part of this project — a chance to work alongside top-tier engineers who are pushing the industrial sector forward.
To make the project even more effective and ensure higher solution quality, I would have brought a QA specialist on board from the very beginning. In reality, communication was limited to the development team only.
Next case